Thyroid Disease Prediction Using Two Tier Ensemble Classifier

Abstract

Data mining algorithms offer clean way to clear up hassle in scientific records analysis. Data mining supports in complicated information evaluation to discover each issue in dataset. Now-a-days every human being suffers a terrific health. The existing style of everyone is very speedy so it’s very hard to hold his fitness. Every human being can't effortlessly keep the hormone levels within the normal frame. Nowadays hormone disturbance is an essential issue in humans. The fundamental issue that troubles behind thyroid diseases is hormonal disturbance. Thyroid disease is the severe endocrine disease that adversely affects the wellbeing of the human. In such case, disease prediction model can play an important role in disease prognosis. The work deploys wrapper method for feature selection. We have identified top 5 machine learning algorithms for thyroid prediction and these five classifiers are consolidated in the gathering of three with heterogeneous permutations to further improve the performance of deployed classifiers. Majority voting scheme is deployed on the ensemble methods for getting results. The proposed model accomplishes the accurateness of 99.05 % with 67:33 proportions of training and test set. The outcome uncovered that the accuracy of proposed 2-tier ensemble model is increased as compared with single machine learning classifiers.